Buffalo Crash

[Tom Goodrick]

A Dash 8 crashed into a house and burned during an approach to Buffalo International Airport last night (10:30 PM EST). All 49 people on board plus one on the ground were killed. The area just east of Niagra Falls was experiencing rain and snow and many aircraft flying the same approach reported icing. The aircraft was nearly new, delivered only last May. The crew was fairly experienced. It was near the outer marker. I saw one main landing gear sticking out from the wing in a video of the wreckage that seemed to have been deployed.

This time of year, icing conditions on approach to Buffalo are normal. The same type and probably the same exact aircraft and crew have made such approaches many times in recent months. Buffalo is famous for getting "lake-effect" snow almost daily as wind blows moisture off the Great Lakes near it.

The effects of ice on any aircraft are numerous and varied, generally to be avoided. Ice builds up on the wings, tail and props distorting their aerodynamic performance. lift is degraded, drag is increased and controls become less effective. Ice can block pitot tube ports taking out airspeed and altitude. I am not sure how much a problem it is but I would thing ice build -up could reduce the effectiveness of radar and radio antennas so even radar altimeters or radio altimeters could be reduced in accuracy. VOR accuracy would not deteriorate too much but ADF and ILS accuracy could be reduced. Then there is the effect simply of weight distributed differently. Extended gear could collect enough ice to increase the drag considerably. On turboprops the ice shed from the prop hub and the inlet edges can go iinto the engines causing them to fail.

in FS9 all we ever see of icing effects is the loss of airspeed when the pitot tube ices over. That problem is easily fixed by throwing a switch.

[Tom Goodrick]

Today, after a preliminary review of the flight data recorder for the crashed aircraft, an NTSB spokesman reported that the loss of control began as the gear and flaps were being lowered. This seems to indicate that the buildup of ice effects was not noticeable until the changes in drag, pitch and lift occurred as flaps and gear were lowered. Then the aircraft stalled with uneven forces and it was already too late. Anti-ice features had been turned on but did not have the expected effect.

[flaminghotsauce]

I think I read somewhere that the pilots were well aware of the icing, as it was on the windshield.

[Tom Goodrick]

In an unusual step, the NTSB has been giving out detailed info at press conferences about what has been learned from the cockpit voice recorder. Yes, they had been aware of it for some time. The question is whether the boots they had were adequate for the icing encountered.

It will be interesting to find out what the data recorder says. Evidence at the site and witness descriptions indicate the plane was stalled and in a flat spin as it came down directly on the house. There was very little direct horizontal dispersion, only indirect dispersion from the explosion. Experienced pilots have said that ice is seldom perfectly symetrical. When the flaps were deployed, the roll disturbance began indicating the flaps contributed adversely with the ice that had built up. It is also possible the ice on one wing interferred with flap deployment.

The crew were discussing going around.

[hanspetter]

The main question is what can be done to avoid or circumvent this problem? Should Buffalo be closed on days like this and the flight be directed to another airport? Severe icing is more common in GA aircraft since they tend to spend more time in the clouds and may be likely to have poorer flight plans. A scheduled commercial flight should be better prepared especially since lake effect snow is standard during the Buffalo winter. Who should have stopped this before it went too far?

[Tom Goodrick]

Those questions are hard to answer. The full NTSB report must be considered to find the exact cause, if that is possible. The obvious difficulty is that the ice evidence dissappeared a second or two after the crash. Still the flight data recorder should tell in what exact way performance was hindered by the ice.

Consider that a dozen airplanes made successful landings the same night using the same route including aircraft of exactly the same type. They may have been more careful about turning on their de-icing and staying high until the last minute. This condition is prevalent there every year and yet this is the first accident of its type at that airport.

What you say about GA is not true simply because very few GA aircraft are cleared for flight in known icing. Those pilots are staying home. You cannot fly an aircraft in the US without first obtaining a weather briefing dealing with your flight plan. If the briefer says there is icing at certain places or altitudes, you avoid them or, if they canoot be avoided, you do not fly. Yet icing is sometimes encountered. In that case you try to get out of it as soon as possible. first you might need to climb or descend to an altitude that has been reported clear of ice. Serious icing normally occurs at low levels. You can climb through it or descend quickly into it and land. But doing the latter in an airplane not equipped for known ice would be illegal and you would face punishment (suspension of your license for a period of time.

Many US pilots living in ice-prone regions just do not fly much in the winter. You get a good day on a weekend and they all come out and fly around the airport creating considerable traffic.

The old saying about pilots applies: "There are old pilots and there are bold pilots but there are no old, bold pilots."

[flaminghotsauce]

I watched a video of a discussion of horizontal stabilizer icing this afternoon. It looks as if it had been produced after this crash. It's eerie. Sounds like the pilots were still using the autopilot when they deployed flaps. According to the video I watched, deploying flaps caused the severe reaction and having not been hand flying the aircraft, they were unaware of how far the icing had progressed. The autopilot merely compensated. This must have completely surprised them.

The recovery from a normal wing stall is completely opposite the recovery from the horizontal stabilizer stall. IF they tried to hand fly it they may have done exactly the wrong thing to recover. But I read the autopilot was still engaged.

[hanspetter]

Tom, I'm glad to hear that GA pilots don't mess with icing conditions. I just assumed that the less regulated part of the real flight community would be more likely to run into unexpected problems and be less capable of handling them.

Flaming, how exactly do you recover from a horizontal stabilizer stall?

[Tom Goodrick]

My understanding of stabilizer stalls is:
1) They seldom occur in conventional designs because the stabilizer flies at a lower angle of attack than the wing due to the downwash from the wing.
2) When they do occur, as in the stall of a canaard, it is fatal unless luck intervenes.

Perhaps max power can improve your luck.

This seems to be the case of a flat spin after stall/spin.

I just did a data-base search on icing at the NTSB site getting all reports of crashes where icing was a factor between January '98 and December '07, a period when I expected all reports to be final. There were 79 pages with 10 reports per page on 78 of them. So you could say we get about 78 incidents per year on average. (Not all are fatal.) I read the PDF report on the latest one. It was for a Cessna 310R equipped with de-icing boots on wings and tail surfaces. It was flown by an experienced pilot with a commercial rating. He carried one passenger. The passenger died when the plane fell out of the sky in icing conditions just past the outer marker on an ILS approach in northern Michigan.

This guy had reports through ATC of icing below 5,000 ft in clouds down to 2100 ft msl. But it was near the end of a long uneventful flight and he had de-icing equipment. He had just reported his position on the ILS at the outer marker a minute before radar contact was lost. The final descent was extremely steep leaving part of a wing in the top of a tree. At the stall he should have been out of the clouds and done with the icing.

[flaminghotsauce]

Actually, more power is the exact wrong thing to do. Everything you do to recover from a wing stall you do the opposite.

If in fact they did tail stall the airplane in ice, they probably immediately powered up and nosed down, which is what the tail stall is doing already. Once the initial reaction is wrong, there's not enough altitude to recover.

It's so rare, however, one must be regularly trained to notice the oncoming effects to even recognize that it's not a normal wing stall. The stick gets soft forward, even pulling itself forward. This is why one must not push forward like a normal stall. Increasing power exacerbates the effects, increases the force necessary to keep the nose up, so powering up will make it worse also.

It's a bad deal when you're not expecting it. It goes against everything we as pilots are trained to do at stall.

[Tom Goodrick]

Recent releases by the NTSB indicate that improper pilot technique contributed, given the icing situation. The autopilot remained on during the crash. The policy of this company is to disconnect the autopilot when icing is suspected so that a pilot can feel the changes due to ice.

If there was enough ice on the elevator to stall it, I don't think any recovery technique would work.

Flaming, you seem to be saying the best course of action in an elevator stall is to do nothing! Wouldn't that be likely to allow falling backward? It seems full power might at least induce motion toward the nose which could lead to recovery.

[hanspetter]

We're talking T tail in this case. Thus, the tail is above the wings' downwash. It could very well be that the T tail design has some issues regarding icing and recovery.

[flaminghotsauce]

Remember the tail is producing lift, but downward. It's an upside down wing. When it stalls, it noses the plane over. When test pilots were doing the tail stall flights, the stall would come with increased speed.

The proper recovery from a tail stall is to decrease speed and pull back. I know it seems wrong, but the tail wing is upside down.

Why do we push nose down during a normal stall? To decrease the angle of attack. Same thing with more power, to get more forward speed which will reduce the angle of attack as well.

The tail stall is backwards land. The downward lift goes away, the rear of the aircraft goes up, nose goes down. Let me go see if I can find that video.

[flaminghotsauce]

This is it:

www.patricksaviation.com/videos/Contrail_25/3927/

Hopefully the video explains it better than I.

[flaminghotsauce]

More information here:
online.wsj.com/article/SB123492905826906821.html?mod=djemalertNEWS

It seems the plane was entering stall, the pilot nosed UP 31 degrees and shoved the throttles forward. It's more confusing now with more information that it was before.